Salmon only processes human reads from PDX data

Question

[jrleary]

My lab uses Salmon to align / quantify our bulk RNAseq data. It typically runs without a hitch, but recently we've had two patient derived xenograft samples, with a mix of mouse and human DNA, give the message:

Submitting salmon

Run human only on original .fasta

Our PDX runs typically output a separate graft (human) and host (mouse) counts matrix, but for some reason these two samples do not. Has anyone faced this issue before? Any help / tips would be greatly appreciated, as I can't seem to find any mention of this error message in the code on GitHub, on StackOverflow, or here on BioStars.

[sandeep.amberkar18]

@jrleary - While using Salmon for your use case of PDX samples, which reference do you use? Separate references for mouse and human or a combined reference? I have a similar scenario and I could use your help.

Thanks in advance.

Answer 1

[Rob]

Hi jrleary,

I don't have an answer for this, but I can say that Run human only on original .fasta is not an error message generated by salmon itself. Are there some extra checks that happen when you submit your job?

[jrleary]

You are correct, the original author of the pipeline included some esoteric if, then statements involving the number of unambiguous reads compared with the total number of reads before running salmon. Thanks much!